llama.cpp
llama : Support llama 4 text-only
#12791
Merged
Overview
Commits
21
Changes
View On GitHub
Commits
  • llama4 conversion
    ngxson committed 1 year ago
  • initial support, no chat template
    ngxson committed 1 year ago
  • clean up a bit
    ngxson committed 1 year ago
  • fix tokenizer conversion
    ngxson committed 1 year ago
  • correct hparams
    ngxson committed 1 year ago
  • try this
    ngxson committed 1 year ago
  • fix shexp
    ngxson committed 1 year ago
  • ffn_inp_normed
    ngxson committed 1 year ago
  • chat template
    ngxson committed 1 year ago
  • clean up model conversion
    ngxson committed 1 year ago
  • add_bos
    ngxson committed 1 year ago
  • add scale_before_ffn
    ngxson committed 1 year ago
  • fix order
    ngxson committed 1 year ago
  • Merge branch 'master' into xsn/llama4
    ngxson committed 1 year ago
  • weight_before_ffn
    ngxson committed 1 year ago
  • llm_graph_input_attn_temp
    ngxson committed 1 year ago
  • add chunk attn mask
    ngxson committed 1 year ago
  • build_inp_attn_scale()
    ngxson committed 1 year ago
  • add comment about ggml_repeat
    ngxson committed 1 year ago
  • clarify comments
    ngxson committed 1 year ago
  • fix build
    ngxson committed 1 year ago
FAQ Terms Privacy Refunds Impressum
Loading